SDHB-Deficient Cancers: The Role of Mutations That Impair Iron Sulfur Cluster Delivery

Abstract

Background:

Mutations in the Fe-S cluster-containing SDHB subunit of succinate dehydrogenase cause familial cancer syndromes. Recently the tripeptide motif L(I)YR was identified in the Fe-S recipient protein SDHB, to which the cochaperone HSC20 binds.

Methods:

In order to characterize the metabolic basis of SDH-deficient cancers we performed stable isotope-resolved metabolomics in a novel SDHB-deficient renal cell carcinoma cell line and conducted bioinformatics and biochemical screening to analyze Fe-S cluster acquisition and assembly of SDH in the presence of other cancer-causing SDHB mutations.

Results:

We found that the SDHB^R46Q mutation in UOK269 cells disrupted binding of HSC20, causing rapid degradation of SDHB. In the absence of SDHB, respiration was undetectable in UOK269 cells, succinate was elevated to 351.4±63.2 nmol/mg cellular protein, and glutamine became the main source of TCA cycle metabolites through reductive carboxylation. Furthermore, HIF1α, but not HIF2α, increased markedly and the cells showed a strong DNA CpG island methylator phenotype (CIMP). Biochemical and bioinformatic screening revealed that 37% of disease-causing missense mutations in SDHB were located in either the L(I)YR Fe-S transfer motifs or in the 11 Fe-S cluster-ligating cysteines.

Conclusions:

These findings provide a conceptual framework for understanding how particular mutations disproportionately cause the loss of SDH activity, resulting in accumulation of succinate and metabolic remodeling in SDHB cancer syndromes.

Understanding the altered metabolism of cancer cells is crucial for the development of effective forms of therapy for patients affected by this disease. In the 1920s, Otto Warburg demonstrated that many cancers depend on glycolysis rather than respiration for energy production (the Warburg effect), even in the presence of oxygen (1). Mutations in two citric acid cycle enzymes, fumarate hydratase (FH) and succinate dehydrogenase (SDH), cause familial cancer syndromes that are prototypic examples of the Warburg effect in cancer (2). While patients with germline FH mutations are at risk for the development of cutaneous and uterine leiomyomas and an aggressive form of type 2 papillary kidney cancer characterized by a metabolic shift to aerobic glycolysis and glutamine-dependent reductive carboxylation (3,4), those with germline mutations in succinate dehydrogenase are at risk for the development of paragangliomas, pheochromocytomas, gastrointestinal stromal tumors (GIST), as well as an aggressive form of oncocytic kidney cancer (SDH-RCC) (5–8).

Succinate dehydrogenase, which functions as complex II in the mitochondrial respiratory chain, is a complex made up of SDHA, SDHB, SDHC, and SDHD subunits. SDHA couples the oxidation of succinate to fumarate with the reduction of covalently bound FAD+ to FADH₂. Three iron-sulfur (Fe-S) clusters in SDHB facilitate transfer of electrons from FADH₂ to ubiquinone, which is bound via the membrane-embedded SDHC and SDHD subunits (9). We recently showed that succinate dehydrogenase assembly and function are dependent on two highly conserved L(I)YR motifs in SDHB, which confer critical specificity for iron sulfur cluster delivery. The pathogenic mutation SDHB^R46Q alters the first L(I)YR motif by changing IYR to IYQ and causes impaired Fe-S cluster incorporation into SDHB, thus rendering the protein unstable (10).

Here we report the characterization of an SDHB-deficient renal cell carcinoma cell line from a young patient carrying the SDHB^R46Q mutation, which was used to explore the altered metabolism of SDH-deficient cancers and gain mechanistic insights into the delivery of Fe-S clusters to SDHB. Metabolic profiling demonstrated a metabolic shift to aerobic glycolysis as well as dependence on reductive carboxylation of glutamine-derived carbon in the TCA cycle. Finally, a systematic biochemical and bioinformatic analysis of reported SDHB cancer-causing missense mutations in neuroendocrine and renal tumors revealed that residues involved in acquisition or ligation of Fe-S clusters accounted for a high percentage of SDHB-related tumors.

Methods

Patient Characteristics

The patient, who was evaluated at the National Institutes of Health on a Urologic Oncology Branch, National Cancer Institute (NCI) protocol approved by the NCI Institutional Review Board, gave written informed consent for participation in this study. The clinical course and presentation of this patient are described in the Supplementary Materials (available online) and have been described previously (8).

Tissue Culture Procedures

See the Supplementary Materials (available online).

Native PAGE (BN-PAGE) and Immunoblot

The NativePAGE Novex Bis-Tris gel system (Invitrogen, Carlsbad, CA) was used for the analysis of native membrane protein complexes and native mitochondrial matrix complexes, with several modifications, as already described (10). Anti-SDHA and SDHB antibodies were from Mitosciences (Eugene, OR), and rabbit anti-Tom20 was from Santa Cruz Biotechnology (Santa Cruz, CA).

In-Gel and Spectrophotometric Complex II Activities

Detailed protocols can be found in the Supplementary Methods (available online). Complex II (SQR) activity in whole-cell extracts was measured using a microplate assay from Abcam (Cambridge, UK).

Seahorse and Metabolic Tracer Analysis

See the Supplementary Materials (available online).

Statistical Analysis

Statistical analysis was performed using parametric unpaired, two-tailed t tests with 99% confidence intervals, and P values of less than .05 were considered statistically significant. All error bars presented in this work represent standard deviation.

Results

UOK269: A Renal Cell Carcinoma Cell Line Characterized by an R46Q Mutation in the IYR Motif of SDHB

Mutations in mitochondrial complex II genes SDHB, SDHC, and SDHD have recently been found to cause the familial kidney cancers, which are characterized by an early onset of disease and highly aggressive growth (6,8). A radical nephrectomy was performed on a woman age thirty-two years to remove a 5.2cm T3aN0M1 renal tumor (Figure 1A) (see the Supplementary Materials, available online), which showed oncocytic morphology (Figure 1B)(8). The patient died of widely metastatic kidney cancer six months after surgery. A kidney cancer cell line, UOK269, was established following surgical excision of the tumor (Figure 1C). Germline sequencing of candidate familial kidney cancer–causing genes identified a heterozygous point mutation (CGA to CAA) in the SDHB gene, resulting in a codon change at the conserved amino acid position 46 (R46Q) (Figure 1D). Genetic testing of the patient’s family members revealed that the patient’s father and paternal aunt, the latter of whom had a history of paraganglioma, carried the R46Q mutation (8). The patient’s tumor and tumor-derived UOK269 cell line showed loss of the remaining wild-type SDHB allele and retention of only the SDHB mutated allele (Figure 1D). Importantly, the R46Q mutation occurs in the first L(I)YR motif of SDHB, which was recently shown to be critical for acquisition of Fe-S clusters in SDHB and for subsequent incorporation of SDHB into mitochondrial complex II (10).

Figure 1.

An SDHB-deficient tumor cell line was derived from culture of a renal cell carcinoma with an R46Q mutation in SDHB. A) Contrast-enhanced computed tomography (CT) scan revealed a 5.2cm mass on the left kidney (white arrow). B) The patient underwent a radical nephrectomy and pathologic examination revealed that the tumor had oncocytic morphology, as described previously (8). Scale bar = 50 µm. C) A cultured explant from the tumor yielded the continuous cell line, UOK269. Scale bar = 10 µm. D) Genomic DNA sequencing revealed that the patient harbored a heterozygous germ-line point mutation (R46Q) in the SDHB gene and that both the tumor and the tumor-derived cell line UOK269 showed loss of heterozygosity at this allele, whereas both the normal G residue and the abnormal A transition were detectable in somatic DNA derived from the patient’s blood. E) Western blot analyses of extracts made from normal kidneys and from this patient’s tumor (R46Q) as well as an unrelated tumor (R90*) also harboring a premature stop codon in sdhb revealed that both tumors showed a profound loss of SDHB protein, whereas the complex II subunit, SDHA, was still present in the tumors. The ponceau S-stained membrane served as a loading control for assessment of total protein levels. F) CpG island methylation status was compared between DNA extracts of UOK269 and HRCE cells using the Illumina HumanMethylation450 BeadChip array. The number of distinct CpG island probes showing increased differential methylation in UOK269 (calculated as the UOK269 β-value minus the HRCE β-value) of 0.5 or higher, indicating hypermethylation, are shown graphed in intervals of 0.1. G) Comparison of the 17 004 CpG island probes hypermethylated in UOK269 vs normal kidney renal cortex tissue was visualized using cluster analysis. H&E = hematoxylin and eosin.

To evaluate expression of SDHB^R46Q in the SDH-RCC patient’s tumor, western blot analyses were performed on tumor extracts derived from the patient harboring the R46Q mutation, as well as from the tumor of a patient who harbored a germline SDHB nonsense mutation, which caused premature truncation at amino acid residue R90, SDHB^R90* (Figure 1E). Both tumors showed dramatic reduction of SDHB protein levels as compared with normal kidney tissue, despite ample expression of the separately encoded SDHA subunit of respiratory complex II (Figure 1E).

Tumors with SDHx mutations have recently been shown to have a global DNA CpG island methylator phenotype (CIMP) attributed to the product-level inhibition of demethylases such as TET2 by elevated succinate (11,12). The methylation status of the UOK269 cells was evaluated using the Illumina HumanMethylation450 BeadChip array and compared with cultured primary human renal cortical epithelial (HRCE) cells. Differential methylation values for each of the approximately 150 000 Illumina probes positioned within CpG islands were calculated by subtracting the β-values of the control probe (HRCE) from β-values of the UOK269 probe. A positive differential methylation value of 0.5 or highger was considered to represent hypermethylation and identified 17 004 distinct CpG island probes in UOK269 (11%) (Figure 1F). Conversely, a negative differential methylation value of -0.5 or less was considered to represent hypomethylation and identified only 135 distinct CpG island probes in UOK269. Increased methylation was also observed in regions outside of CpG islands. These results support the presence of a hypermethylation phenotype in UOK269 cells. Furthermore, cluster analysis of the 17 004 hypermethylated CpG island probes in UOK269, performed by comparing UOK269 to two normal kidney renal cortex tissue DNA samples obtained from the TCGA clear cell RCC analysis (13), suggested that DNA hypermethylation in UOK269 cells may be an aberrant phenotype related to SDH deficiency, as this pattern was not present in the normal tissue (Figure 1G).

Evaluation of Metabolic Parameters and HIF1α Expression in UOK269 Cells

Loss of enzymatic components of the TCA cycle is expected to impair oxidative phosphorylation, and decreased oxygen consumption has been observed in tumor-derived cells harboring mutations in the gene for the TCA cycle enzyme fumarate hydratase (14). Analysis of complex II activity demonstrated a profound lack of succinate dehydrogenase (SDH) activity in SDHB-deficient UOK269 cells (mean±SD; 0.25±0.04 decrease in absorbance at 600nm per minute per mg cellular protein; UOK269 vs HEK293 P < .01), as compared with HEK293 cells (mean±SD; 14.16±1.62 decrease in absorbance at 600nm per minute per mg cellular protein) or UOK269 cells engineered to stably express wild-type SDHB gene by lentiviral transduction (UOK269WT; mean±SD 10.13±2.02 decrease in absorbance at 600nm per minute per mg cellular protein; UOK269WT vs HEK293 P > .05) (Figure 2A). Complex II activity was not restored in UOK269 cells transfected with the backbone vector lacking the SDHB gene (UOK269EV; mean±SD; 0.39±0.24 decrease in absorbance at 600nm per minute per mg cellular protein; UOK269EV vs HEK293 P < .01) (Figure 2A). Oxygen consumption, measured by Seahorse Flux Analyzer, in SDHB-deficient UOK269 cells was below the limits of detection (Figure 2B). UOK269 cells showed strong, glucose-stimulated acidification of the culture medium (Figure 2C), suggesting that these cells rely heavily on glucose metabolism and lactic acid production for growth and cell division. In contrast, UOK269WT and HEK293 cells, which have intact mitochondrial complex II, showed increased respiration (oxygen consumption rate) and lower glycolytic activity (extracellular acidification rate) (Figure 2D).

Figure 2.

Measurement of complex II activity, oxygen consumption, extracellular acidification, and HIF1α expression in UOK269 cells. A) Measurement of complex II activity (succinate→ubiquinone) revealed that UOK269 cells and UOK269EV cells transfected with an empty vector showed dramatically reduced complex II activity levels, whereas UOK269 cells stably transfected with a recombinant clone of the SDHB gene (UOK269WT) showed restoration of complex II activity. Error bars are standard deviation. **P < .01. B) Analysis of oxygen consumption rate using the Seahorse Flux Analyzer revealed that oxygen consumption rates (OCR) in UOK269 and UOK269EV were undetectable, whereas SDHB-restored UOK269WT cells and HEK293 cells both showed robust OCR. Error bars are standard deviation. C) UOK269 and UOK269EV cells showed a high rate of extracellular acidification (ECAR), which was stimulated by addition of glucose to the assay medium. D) The OCR was plotted against the ECAR to demonstrate that, in stark contrast to HEK293A and UOK269WT cells, extracellular acidification in UOK269 and UOK269EV was not associated with a measurable rate of oxygen consumption during the experiment. FH-deficient UOK262 cells showed low but measurable OCR under the same conditions. E) Western blots of nuclear extracts demonstrated marked HIF1α in UOK269 cells compared with HEK293 cells and human MCH46 fibroblasts. However, expression of wild-type SDHB in UOK269 cells (UOK269WT) reversed the induction of HIF1α expression, relating the induction of HIF1α to loss of SDHB. The iron chelator desferrioxamine (DFO) also induced HIF1α expression in fibroblasts. F) In contrast, HIF2α protein levels were not elevated in UOK269 as compared with UOK269WT cells or HEK293 cells. MCH46 fibroblasts exposed to hypoxia for six hours served as the control for induction of HIF2α.

Recently, there has been growing appreciation of the ability of specific cellular metabolites to act as pro-oncogenic factors that can stimulate carcinogenesis (2). Succinate is one such ‘oncometabolite’ because elevated cellular succinate can cause inhibition of a class of iron- and α-ketoglutarate-dependent dioxygenases, the prolyl hydroxylases (PHDs), which are responsible for targeting hypoxia inducible factors HIF1α and HIF2α for degradation (15). Consistently, we found greatly elevated HIF1α protein levels in nuclear extracts from SDHB-deficient UOK269 cells as compared with HEK293 cells and UOK269WT cells, in which SDH activity was restored (Figure 2E). The human fibroblast cell line MCH46 depleted of iron by treatment with the chelator desferrioxamine (DFO) served as a positive control for identification of elevated HIF1α because of stabilization of the protein through inhibition of the PHDs enzymatic reaction (Figure 2E). HIF2α expression in nuclear extracts from UOK269 cells did not change compared with UOK269WT, HEK293, or fibroblast cell lines. MCH46 fibroblasts grown in the presence of low oxygen for six hours served as the positive control. Together, these data suggest that elevated HIF1α levels in SDHB-deficient UOK269 cells might be caused by inhibition of the PHDs-mediated degradation of HIF1α because of high intracellular succinate.

Sources of TCA Cycle Intermediates in SDHB-Deficient UOK269 Cells

Tumors rely on glucose and glutamine for energy production and biosynthesis of the macromolecules necessary for continuous proliferation (16). Given the pivotal role of SDH as the interface between the TCA cycle and the mitochondrial respiratory chain, we evaluated TCA cycle metabolite levels using stable isotope-resolved metabolomics. We observed dramatically high levels of intracellular succinate in UOK269 cells (n = 3, mean±SD, 351.4±63.2 nmol/mg protein) (Figure 3A), as well as secretion of succinate into the culture medium (Figure 3B). By culturing UOK269 cells with either ¹³C₆-glucose or ¹³C₅,¹⁵N₂-glutamine, we found that the secreted extracellular succinate was derived from ¹³C₅,¹⁵N₂-glutamine (Figure 3B), whereas extracellular lactate was produced exclusively from ¹³C₆-glucose (Figure 3C). These findings demonstrate that glucose is preferentially directed toward lactic acid fermentation and secretion, whereas secreted succinate is derived from glutamine in SDHB-deficient UOK269 cells.

Figure 3.

Metabolic analysis of SDHB-deficient UOK269 cells using Stable Istope-Resolved Metabolomics. A) GCMS analysis of selected metabolite concentrations in UOK269 cells demonstrated that intracellular succinate levels were 351.4±63.2 nmol/mg cellular protein. B) ¹H-(32) HSQC NMR analysis of culture media demonstrated that ¹³C-glutamine-derived succinate was secreted by UOK269 cultures. C) Extracellular lactate secreted by UOK269 cells was derived exclusively from ¹³C-glucose and not from ¹³C-glutamine. D) ¹H-(32) HSQC NMR of UOK269 cell extracts demonstrated that ¹³C-glucose-derived carbons were incorporated into lactate, alanine, nucleotides, and glycogen, whereas the TCA-derived intermediates glutamate, succinate, and glutathione (which contains glutamate) derived their carbons from ¹³C-glutamine. GC-MS analysis of isotopologues of succinate (E), malate (F), fumarate (G), aspartate (H), and citrate (I) in UOK269 cell extracts grown in the presence of ¹³C₆-glucose, ¹³C₅,¹⁵N₂-glutamine, or singly labeled ¹³C₁-glutamine for 24 hours underscored that glutamine was the dominant carbon source for these metabolites. J) Schematic diagram illustrates that glucose was mainly converted to lactate, alanine, nucleotides, and glycogen, whereas glutamine was a dominant carbon source of TCA cycle intermediates.

Next, we utilized 1D-heteronuclear single quantum coherence nuclear magnetic resonance (HSQC NMR) spectroscopy to examine the distribution of intracellular metabolites in SDHB-deficient UOK269 cells following growth for 24 hours in the presence of either ¹³C₆-glucose or ¹³C₅,¹⁵N₂-glutamine (Figure 3D). These data demonstrated that glucose-derived signals were found primarily in the ribose moieties of nucleotides (ie, AXP, UXP, NAD+), as well as in glycogen, lactate, and alanine (Figure 3D). Conversely, while labeling of UOK269 cells with ¹³C₅,¹⁵N₂-glutamine resulted in robust labeling of intracellular succinate, glutamate, and glutathione, it did not result in appreciable labeling of lactate, alanine, or nucleotides (Figure 3D). Analysis of these stable isotope–labeled UOK269 cell extracts using gas chromatography–mass spectrometry (GC-MS) further demonstrated that the highly elevated intracellular succinate in SDHB-deficient UOK269 cells was derived nearly exclusively from glutamine, as incubation of UOK269 cells with ¹³C₅,¹⁵N₂-glutamine resulted in fully labeled (m+4) intracellular succinate. Singly labeled ¹³C-1-glutamine did not label succinate (Figure 3E), indicating that the ¹³C label was lost as carbon dioxide (CO₂) during the α-ketoglutarate dehydrogenase reaction. In contrast, incubation of UOK269 cells with ¹³C-1-glutamine resulted in robust m+1 labeling of malate, aspartate, fumarate, and citrate, indicating that glutamine was likely incorporated into citrate by reductive carboxylation of glutamine-derived α-ketoglutarate through the activity of isocitrate dehydrogenase (Figure 3, F-J; Supplementary Figure 1, available online)(3).

Functional Analysis of Disease-Causing SDHB Mutations

The metabolic analyses described above demonstrate the remarkable ability of SDH-deficient human renal cancer cells to adapt to the loss of a crucial component of the TCA cycle and the mitochondrial respiratory chain. An understanding of SDH-related tumorigenesis is crucial because mutations in succinate dehydrogenase genes, including SDHB, increase susceptibility to development of paragangliomas and pheochromocytomas (17,18), renal cell carcinoma (6,8), and gastrointestinal stromal tumors (19). SDHB is the Fe-S subunit of Complex II and contains two highly conserved L(I)YR motifs that are essential for acquisition of Fe-S clusters by recruiting the Fe-S transfer machinery (10). We searched the Leiden Open Variation Database (LOVD)(20) for germline oncogenic missense mutations in SDHB that affect amino acid residues that are important for Fe-S cluster acquisition (L(I)YR motif amino acid residues) or ligation (cysteines). The 3D structure of porcine SDHB(9) (Figure 4A), which is 96% identical to human SDHB, shows the position of the two L(I)YR motifs and the three Fe-S clusters, which are also highlighted in the primary sequence of human SDHB (Figure 4B). In order to evaluate whether oncogenic mutations clustered near residues involved in acquisition or ligation of the Fe-S clusters, we catalogued the number and position of SDHB missense mutations. Notably, missense mutations of residues in the first (I₄₄Y₄₅R₄₆) and in the second (L₂₄₀Y₂₄₁R₂₄₂) L(I)YR motifs accounted for 12% of reported tumors. Mutations of the cysteinyl ligands of the first, second, and third Fe-S clusters accounted for an additional 8%, 12%, and 5% of tumors, respectively (Figure 4C). Altogether, mutations in the L(I)YR motifs or cysteinyl ligands accounted for 37% of families, whereas other mutations were located more diffusely across the SDHB primary sequence (Figure 4C). The occurrence of renal cell carcinomas (SDH-RCCs) and gastrointestinal stromal tumors (GISTs) is less frequent than that of pheochromocytomas and paragangliomas in patients with SDHB mutations and can be considered a more aggressive phenotype. Analyses of reported missense mutations that affected the L(I)YR motifs or the cysteinyl ligands in patients with SDH-RCC or GIST revealed that 50% of these patients had SDHB mutations in the L(I)YR motif amino acid residues (Supplementary Figure 2, available online).

Figure 4.

Missense mutations in the SDHB gene that cause familial cancer syndromes. A) Ribbon representation of the three-dimensional structure of porcine SDHB (PDB ID: 3SFD, 96% identical to human) and (B) the primary sequence of human SDHB. L(I)YR motifs are shown in cyan. The cysteine residues that coordinate the Fe₂-S₂ cluster are in magenta, and the ligands of the Fe₄-S₄ and of Fe₃-S₄ clusters are in green and yellow, respectively. C) Relative incidence of SDHB missense mutations that cause cancer listed in the LOVD database(43). Bars representing the incidence of mutations affecting the LYR motif amino acid residues or the Fe-S cluster-coordinating cysteines are labeled and shown in colors matching with panels (A) and (B).

L(I)YR Mutations and SDH Complex Assembly and Activity

Analyses of the UOK269 cell line and tumor material harboring the homozygous SDHB^R46Q mutation revealed that SDH activity was lost (cell line) (Figure 2A) and SDHB protein was undetectable (tumor) (Figure 1). Therefore, we used UOK269 cells as an experimental system that lacked SDHB and SDH activity to express and characterize reported SDHB pathogenic mutations of the L(I)YR motif amino acid residues or of the Fe-S cluster–ligating cysteines (Table 1). Expression of SDHB^I44N, which harbors a mutation in the first L(I)YR motif of SDHB and has been reported to cause GIST (21), was unable to restore SDH activity in UOK269 cells (Figure 5A). In contrast, UOK269 cells transfected with wild-type SDHB (+SDHB) (Figure 5A) revealed robust SDH activity. As SDHB acquires its Fe-S clusters before forming a functional complex with SDHA, SDHC, and SDHD (10), we used Blue Native-PAGE (BN-PAGE) analysis and immunoblotting (IB) of mitochondrial extracts to assess the various stages of complex II assembly. We found no evidence of mature complex II (CII) formation in UOK269 cells transfected with SDHB^I44N (Figure 5B), though an assembly intermediate, termed CIIb, was detected, which was previously shown to contain HSC20, ISCU, and HSPA9 (10). Importantly, the second LYR motif in SDHB^I44N was intact. Nevertheless, SDHB^I44N failed to incorporate its Fe-S clusters and did not progress into formation of a complex with SDHA (CIIa). Interestingly, steady-state protein levels of SDHB^I44N were reduced compared with wild-type SDHB or to other mutants (Figure 5C). SDHB^R242C, harboring a substitution of the arginine into cysteine in the second L(I)YR motif of SDHB, was also unable to incorporate into a functional complex or restore SDH activity (Figure 5, A and B). Interestingly, levels of SDHB^R242C protein were higher than those of SDHB^I44N, consistent with the previously reported instability of SDHB mutants harboring mutations in the first IYR motif (Figure 5C) (10).

Table 1.

Missense mutations in the LYR motifs of SDHB or in the cysteines that coordinate the clusters reported in the LOVD database*

Variant - exon	Variant - DNA	Variant - protein	Variant - var_type	PGL	Pheo	RCC	GIST	Site affected by the mutation	References
2	c.131T>A	p.Ile44Asn	Missense	—	—	—	GIST	First motif	Celestino et al., 2012 (21)
2	c.136C>G	p.Arg46Gly	Missense	PGL	Pheo	—	—	First motif	Neumann et al., 2004 (30)
2	c.137G>A	p.Arg46Gln	Missense	PGL	Pheo	RCC	—	First motif	Ricketts et al., 2012 (8)
3	c.277T>C	p.Cys93Arg	Missense	PGL	—	—	—	Ligand of [Fe2-S2]	Lima et al., 2007 (31)
3	c.278G>A	p.Cys93Tyr	Missense	PGL	—	—	—	Ligand of [Fe2-S2]	Cascon et al., 2009 (32)
4	c.292T>C	p.Cys98Arg	Missense	PGL	—	—	—	Ligand of [Fe2-S2]	Benn et al., 2006 (33)
4	c.293G>A	p.Cys98Tyr	Missense	PGL	—	—	—	Ligand of [Fe2-S2]	Benn et al., 2006 (33)
4	c.302G>A	p.Cys101Tyr	Missense	PGL	—	—	—	Ligand of [Fe2-S2]	Neumann et al., 2004 (30)
4	c.338G4A	p.Cys113Tyr	Missense	PGL	Pheo	—	—	Ligand of [Fe2-S2]	Ricketts et al., 2010 (34)
6	c.557G>A	p.Cys186Tyr	Missense	PGL	—	—	—	Ligand of [Fe4-S4]	Lima et al., 2007 (31)
6	c.565T>C	p.Cys189Arg	Missense	PGL	—	—	—	Ligand of [Fe4-S4]	Burnichon et al., 2009 (35)
6	c.566G>T	p.Cys189Phe	Missense	PGL	—	—	—	Ligand of [Fe4-S4]	LOVD
6	c.572G>A	p.Cys191Tyr	Missense	PGL	Pheo	—	—	not a cluster ligand	Goffrini et al., 2009 (36)
6	c.574T>C	p.Cys192Arg	Missense	PGL	—	—	—	Ligand of [Fe4-S4]	Neumann et al., 2004 (30)
6	c.575G>A	p.Cys192Tyr	Missense	PGL	—	—	—	Ligand of [Fe4-S4]	Benn et al., 2006 (33)
6	c.587G>A	p.Cys196Tyr	Missense	PGL	—	—	—	Ligand of [3Fe-4S]	Brouwers et al., 2006 (37)
6	c.588C>G	p.Cys196Trp	Missense	—	Pheo	—	—	Ligand of [3Fe-4S]	LOVD
7	c.724C>A	p.Arg242Ser	Missense	PGL	—	—	—	Second motif	Neumann et al., 2009 (38)
7	c.724C>T	p.Arg242Cys	Missense	PGL	—	—	—	Second motif	Schiavi et al., 2006 (39)
7	c.725G>A	p.Arg242His	Missense	PGL	Pheo	—	—	Second motif	Cascon et al., 2009 (40)
7	c.727T>A	p.Cys243Ser	Missense	PGL	—	—	—	Ligand of [Fe3-S4]	Korpershoek et al., 2007 (41)
7	c.745T>G	p.Cys249Gly	Missense	PGL	—	—	—	Ligand of [Fe3-S4]	Burnichon et al., 2009 (35)
7	c.746G>A	p.Cys249Tyr	Missense	PGL	—	—	—	Ligand of [Fe3-S4]	Burnichon et al., 2009 (35)
7	c.758G>A	p.Cys253Tyr	Missense	—	Pheo	—	—	Ligand of [Fe4-S4]	Amar et al., 2007 (42)

* GIST = gastrointestinal stromal tumor; LOVD = Leiden Open Variation Database; PGL = paraganglioma; Pheo = pheochromocytoma; RCC = renal cell carcinoma.

Figure 5.

Expression of SDHB clones with missense mutations in either of the two LYR motifs in SDHB null UOK269 cells. A) In-gel succinate dehydrogenase (SDH) activity of complex II (CII), (B) native immunoblot to SDHB, and (C) western blots with antibodies to SDHA, SDHB, and HSPA9 on mitochondrial extracts from HEK293, UOK269, UOK269WT, and UOK269 cells transfected for 48 hours with pCMV-SDHB-F/M (expressing FLAG- tagged SDHB wild-type) or with SDHB constructs bearing mutations in either the two L(I)YR motifs, SDHB^I44N (of the IYR motif) or SDHB^R242C of the second LYR motif (+SDHB^I44N and +SDHB^R242C). In (B), CIIa and CIIb are complex II assembly intermediates (a- c, n = 5 biological replicates).

Cysteine Mutations and SDH Complex Assembly and Activity

We also tested the effects on SDH assembly and activity of mutations of the cysteines that ligate each of the three Fe-S clusters (Figure 6A). None of the SDHB mutants harboring cancer-causing mutations in the Fe-S cluster–ligating cysteines exhibited any succinate-coenzyme Q reductase (SQR) activity (Figure 6A). The native immunoblots for SDHB protein also demonstrated that none of the SDHB mutants progressed to association with SDHA (complex CIIa) or to full SDH complex formation (CII) (Figure 6B). Nevertheless, the transfected missense proteins were detectable by immunoblot (Figure 6C). Finally, quantitative spectrophotometric assay of SDH activity in mitochondrial extracts from UOK269 cells expressing the SDHB cancer-causing mutations in L(I)YR motifs or in the cysteines revealed that none of the mutants was able to restore SDH activity (Figures 6D; Supplementary Figure 1, available online). Our biochemical and functional characterization of cancer-causing missense mutations affecting the L(I)YR motifs or the Fe-S cluster cysteinyl ligands of SDHB demonstrated the detrimental effects of mutations, which occur in positions that are crucial for biogenesis of SDHB and assembly of the SDH complex.

Figure 6.

Expression of SDHB clones bearing missense mutations in the Fe-S cluster- coordinating cysteines of SDHB in SDHB-null UOK269 cells. A) In-gel succinate co-Q oxidoreductase (SQR) activity of complex II (CII) upon transfection with SDHB clones containing mutations corresponding to those of reported in cancer patients, (B) native immunoblot to SDHB, demonstrating failure of each mutation to reconstitute functional SDH, and (C) western blots with antibodies to SDHA, SDHB, and TOM20 on mitochondrial extracts from HEK293, UOK269, UOK269WT, and UOK269 cells transfected for 48 hours with pCMV-SDHB-F/M (expressing FLAG- tagged SDHB wild-type) or with the mutants of the L(I)YR motifs and of the cluster- coordinating cysteines. Cysteines 98 and 113 (+SDHB^C98R and +SDHB^C113Y) were replaced by arginine and tyrosine, respectively; cys189 and cys253 (+SDHB^C189R and +SDHB^C253Y) by Arg and Tyr, and the Fe₃-S₄ cluster-coordinating cysteines cys243 and cys249 (+SDHB^C243S and +SDHB^C249Y) were replaced by Ser and Tyr, respectively. In (B), CIIa and CIIb are complex II assembly intermediates. D) Spectrophotometric measurements of succinate dehydrogenase (SDH) activity performed on 10 μg of mitochondrial extracts in a reaction volume of 200 μl. Results are expressed as changes in the absorbance per minute (ΔA/min). Considering that the molar absorbance of the INT-formazan at 500nm is 19 300, the catalytic activity can be measured as: U/ml= ΔA/min x 5,18. Error bars indicate standard deviation (a- d, n = 5 biological replicates).

Discussion

Alteration of cellular metabolism is a hallmark of cancer, contributing to its initiation, growth, and progression (22,23). Recent studies have shown that renal cancer is associated with mutations of specific genes that affect cellular metabolism (23,24). Clear examples of this are the hereditary cancer syndromes characterized by germline mutation of the genes for the TCA cycle enzymes fumarate hydratase (FH) and succinate dehydrogenase (SDH) (6,25). In the present study, we evaluated the role of cancer-causing SDHB mutations that impair the delivery of Fe-S clusters into the SDHB protein. We developed an SDHB-deficient cell line derived from a kidney tumor that was surgically removed from a patient carrying a germline SDHB^R46Q mutation that lies in the first L(I)YR motif required for incorporation of the Fe₂-S₂ cluster into SDHB. The pronounced decrease in SDHB activity in these tumor-derived SDHB^R46Q cells led to a metabolic shift characterized by aerobic glycolysis, a total absence of detectable mitochondrial respiration, as well as markedly elevated HIF1α protein levels, all of which were restored with re-expression of SDHB. These results suggest that, similar to our findings in FH-deficient kidney cancer (4), HIF1α is the predominant HIF in SDH-deficient kidney cancer (and that HIF2α protein levels are unaffected by the elevated succinate levels in UOK269 cells).

We have previously shown that FH-deficient kidney cancer is characterized by glutamine-dependent reductive carboxylation, which is critical for increased fatty acid production needed for rapid cellular growth (3). In the present study we utilized stable isotope–resolved metabolomics to examine the metabolic effects of loss of SDHB in UOK269 kidney cancer cells and found substantial increases in both intracellular as well as extracellular succinate, which is consistent with the phenotype of SDHx-mutated paragangliomas and pheochromocytomas (26,27). In addition, SDHB-deficient renal cell carcinoma was also characterized by glutamine-dependent reductive carboxylation. M+3 labeling of malate, fumarate, and aspartate in UOK269 cells labeled with ¹³C₅,¹⁵N₂-glutamine most likely resulted from cleavage of glutamine-derived ¹³C₅-citrate by ATP citrate lyase in the cytosol, yielding ¹³C₃-oxaloacetate that could, in turn, be converted to other TCA cycle intermediates (Figure 3J; Supplementary Figure 1, available online). In contrast, ¹³C₆-glucose-derived M+3-labeled TCA cycle metabolites likely resulted from mitochondrial pyruvate carboxylase activity in UOK269 cells (Figure 3J).

Nascent Fe-S clusters are assembled on the main scaffold protein ISCU and subsequently transferred to recipient proteins such as SDHB by a dedicated chaperone-cochaperone system consisting of the HSPA9-HSC20 pair (4). We recently found that the cochaperone HSC20 facilitates delivery of Fe-S clusters to specific recipients, including SDHB, by directly binding the L(I)YR tripeptide motif present in Fe-S proteins or in accessory factors that assist biogenesis of Fe-S proteins (10). The two L(I)YR motifs in SDHB engage the HSC20-HSPA9-holo-ISCU complex to facilitate incorporation of three Fe-S clusters into complex II (10). Mutations of the two L(I)YR motifs, or the eleven Fe-S cluster cysteinyl ligands, were frequent causes of SDHB-related cancers, accounting for 37% of the tumors thus far described in association with SDHB mutations. The biochemical characterization of the SDHB mutants harboring disease-causing mutations in either of the two L(I)YR sequences or in the eleven Fe-S cluster–ligating cysteines revealed that the mutations impaired biogenesis of SDHB by preventing acquisition of its prosthetic groups. Because the Fe-S clusters are crucial for electron transport and function, it is not surprising that these mutations completely abrogate SDH activity.

The prevalence of missense mutations in the two L(I)YR motifs of SDHB in familial paraganglioma/pheochromocytoma/GIST/renal cell carcinoma tumor syndromes underscores the importance of this recently identified motif (10). Although these findings offer important organizing principles for understanding SDHB-dependent tumorigenesis, a number of questions remain, including the extent to which succinate-mediated competitive inhibition of 2-oxoglutarate-dependent dioxygenases, including prolyl hydroxylases, a subset of histone demethylases in the JMJD family, and the ten-eleven translocation (TET) family of 5-methyl cytosine (5 mC) family of hydroxylases, contribute to the SDH-deficient tumor phenotype (28). Potential succinate-induced changes include stabilization of HIF1α (15) and hypermethylation of histones and DNA (11,12). Notably, UOK269 cells show a DNA hypermethylation phenotype similar to that observed in other SDHx tumors (11,12,29) and a marked elevation of HIF1α.

One limitation of the present study is that not all disease-causing SDHB missense mutations were tested for their ability to reconstitute complex II assembly and activity in UOK269 cells. It remains to be determined whether some tumorigenic SDHB mutations, such as those that do not affect either the L(I)YR Fe-S transfer motifs or one of the eleven Fe-S cluster–ligating cysteines of SDHB, could result in some level of residual SDH activity that may be reflected in the clinical phenotype. Further in vitro and in vivo biochemical and metabolic analyses of the familial cancer syndromes associated with deficiency of the enzymes of intermediary metabolism will hopefully lead to the identification of novel functions of TCA cycle metabolites and provide the foundation for the development of therapeutic approaches for these and related cancers.

Funding

This research was supported by the Intramural Research Programs of the National Cancer Institute, Center for Cancer Research, and the Eunice Kennedy Shriver National Institute of Child Health and Human Development, as well as through a National Institutes of Health (NIH) Bench-to-Bedside award made possible by the NIH Office of Rare Diseases Research (ORDR), National Center for Advancing Translational Sciences (NCATS), NIH Office of Intramural Research, and NIH grants 5R01ES022191, 3R01ES022191-04S1, and 1U24DK097215-01A1.

The study funders had no role in the design of the study; the collection, analysis, or interpretation of the data; the writing of the manuscript; nor the decision to submit the manuscript for publication.

The authors acknowledge the outstanding editorial and graphics support by Georgia Shaw. The authors appreciate review of the manuscript by Len Neckers.